2

The effectiveness of physical and organisational ergonomic interventions on low back pain and neck pain: a systematic review

Published as: Driessen MT, Proper KI, van Tulder MW, Anema JR, Bongers PM, van der Beek AJ. The effectiveness of physical and organisational ergonomic interventions on low back pain and neck pain: a systematic review. Occup Environ Med 2010;67:277-285.

Abstract

Objective: Ergonomic interventions (physical and organisational) are used to prevent

or reduce low back pain and neck pain among workers. We conducted a systematic review of randomised controlled trials on the effectiveness of ergonomic interventions.

Methods/Results: A total of 10 randomised controlled trials met the inclusion criteria.

There was low to moderate quality evidence that physical and organisational ergonomic interventions were not more effective than no ergonomic intervention on short and long term low back pain and neck pain incidence/prevalence, and short and long term low back pain intensity. There was low quality evidence that a physical ergonomic intervention was significantly more effective for reducing neck pain intensity in the short term (i.e. curved or flat seat pan chair) and the long term (i.e. arm board) than no ergonomic intervention.

Conclusions: The limited number of randomised controlled trial included make it dif-

ficult to answer our broad research question and the results should be interpreted with care. This review, however, provides a solid overview of the high quality epidemiological evidence on the (usually lack of) effectiveness of ergonomic interventions on low back pain and neck pain.

22

Chapter 2

Introduction Low back pain (LBP) and neck pain (NP) are major health problems in the working population and have considerable consequences for workers, employers, and society.1;2 Prevention of these symptoms is imperative. Prevention of LBP and NP can be categorised into primary, secondary and tertiary prevention. The aim of primary prevention is to prevent the onset of symptoms in a healthy working population, while secondary prevention seeks to aid recovery from early symptoms and reduce the risk of symptom recurrence.3 However, due to the high lifetime prevalences of LBP and NP, it is difficult to discriminate between primary and secondary prevention.4 Tertiary prevention is targeted at reducing and assisting the patient to cope with consequent disabilities.3 Because the development of LBP and NP is assumed to be multifactorial (i.e. individual, psychosocial, and physical risk factors play a role)5;6, preventive strategies vary widely. The common strategy of ergonomic intervention is targeted at occupational risk factors such as lifting, physically heavy work, a static posture, frequent bending and twisting, repetitive work, and exposure to vibration4, and can be divided into individual worker interventions, physical ergonomic interventions, and organisational ergonomic interventions.3 Prevention through individual worker interventions mostly consists of 1) physical exercise programmes to improve strength/work capacity, 2) education, instruction or advice on working methods or lifting techniques, or 3) lumbar support or back belts.6 Systematic reviews have shown that with the exception of exercise programs7-9, none of these strategies are effective in preventing LBP.8-12 Evidence on the effectiveness of training to prevent neck pain is inconclusive.13;14 Prevention through physical ergonomic interventions consists of redesigning the workplace (i.e. providing lifting aids and new equipment, and modifying workstations), while prevention through organisational ergonomic interventions encompasses more changes at the system level (i.e. job rotation, modifications to the production system, and job enlargement).6 Previous reviews have shown that there is insufficient evidence of the effectiveness on LBP prevention of the application of physical or organisational ergonomics.5;15;16 Regarding the effectiveness of physical and organisational ergonomics to prevent neck pain, Brewer et al. (2006) found mixed evidence for the effectiveness of arm supports, alternative keyboards, and rest breaks.14 Boocock et al. (2007) concluded that there was moderate evidence that workstation equipment (mouse and keyboard design) and workstation adjustments were effective (i.e. modified lighting, new workplaces, changed office lay out and new software application led to positive health benefits among video display unit workers with NP. Despite the promising results on video workers, insufficient evidence was found to support the use of ergonomic equipment among manufacturing workers with NP.13 In recent years, several randomised controlled trials (RCTs) on the effectiveness of physical and organisational ergonomics on LBP and NP have been conducted and so an up to date systematic review seems warranted.

Chapter 2

23

The objective of this systematic review is to investigate the effectiveness of ergonomic interventions (physical and organisational) in reducing the incidence/prevalence and intensity of LBP and NP among non sick listed workers.

Methods Search strategy

With the help of an experienced librarian, the medical electronic databases Pubmed, EMBASE, PsychINFO and the Cochrane Central Register of Controlled Trials, and the database of the Cochrane Occupation Health Field between 1988 and September 2008 were searched. The sensitive search for RCTs and the search terms for LBP and NP used terms recommended by the Cochrane Back Review Group for searching Pubmed and EMBASE.17 Search strategies in other databases were as close to the sensitive strategy as possible. Verbeek et al. found that no single search term was available to adequately locate occupational health intervention studies.18 Because the terms of ergonomic interventions also vary largely, no search term for ergonomic interventions was added to the search. ‘Musculoskeletal disorders’ was included as search term as this term may incorporate LBP and NP. Because ‘intensity of discomfort’ is frequently used to assess the prevalence of LBP and NP, the term was also added to the search. Two reviewers (MTD and KIP) independently screened the obtained titles and abstracts for eligibility. Studies were eligible when all four inclusion criteria (see below) were met. Inclusion criteria

Inclusion criteria were as follows: - The study was an RCT; - The study population involved a non-sick listed working population; - The intervention met the definition of a physical or organisational ergonomic interven tion, that is: the intervention is targeted on changing the biomechanical exposure at the workplace or on changing the work organisation; - The outcome measure included non-specific LBP or NP incidence/prevalence or intensity of pain. Studies on neck/shoulder pain were considered as NP studies. Exclusion criteria

The exclusion criterion was as follows: - Individual worker interventions. When inclusion or exclusion of a study could not be decided on reading the title and abstract, the full article was retrieved and checked for inclusion. A consensus meeting with a third reviewer (AJvdB) was arranged to sort out disagreements between both reviewers.

24

Chapter 2

Finally, the reference lists of eligible RCTs and relevant review studies were checked for relevant citations. Risk of bias assessment

Using the 12 criteria list of the Cochrane Back Review Group, two reviewers (MTD and KIP) independently assessed the risk of bias of the included RCTs.17 The list and the operationalisation of the criteria are described elsewhere.17 The criteria were scored as ‘yes/no/ don’t know’. If necessary, a consensus meeting with a third reviewer (AJvdB) was arranged to sort out disagreements between the first two reviewers. Subsequently, results of the risk of bias assessment were sent to all first authors and they were asked to provide additional information on the criteria scored as ‘don’t know’. The first authors were also asked to provide additional information on positive or negative scores they disagreed with. RCTs were considered as having ‘low risk of bias’ when at least 50% (six) of the 12 criteria were met, otherwise they were considered as having a ‘high risk of bias’.17 Data extraction

One reviewer (MTD) extracted the data by using a standardised data extraction form.17 Information on study design, randomisation level, population, follow-up period, measurement tools, statistical analyses, outcomes, and effect sizes was extracted. The second reviewer (KIP) checked all data extracted. In case of disagreements, a third reviewer (AJvdB) was consulted. If data were missing, first authors of the studies were contacted and additional information was requested. Data analysis and the GRADE approach

A meta-analysis was performed among studies that reported on the same outcome and had a similar duration of follow-up, that is, short term (closest to 6 months) or long term (closest to 12 months). For studies with a follow-up period of more than 12 months, the final measurement was used in the meta-analysis. If studies compared more than one ergonomic intervention with a control, each ergonomic intervention was analysed separately. To avoid double-counting of studies, only the effects of the ergonomic intervention with the largest effect size were included in the meta-analysis. For comparisons of dichotomous data (eg, incidence/prevalence), if not provided, risk ratios (RR) with a 95% CI were calculated. For comparisons of continuous data (e.g. pain intensity) standardised mean differences with a 95% CI were calculated. The random effects model was used. All analyses were conducted using the RevMan 5 software. The GRADE approach was used to classify the overall quality of the evidence.19;20 For each specific outcome the quality of the evidence was based on five factors: 1) limitations of the study referring to the risk of bias for the results across all studies that measure that specific outcome, 2) consistency of results, 3) directness (generalisability), 4) precision

Chapter 2

25

(sufficient data), and 5) the potential for publication bias. The overall quality of evidence was considered to be high if multiple RCTs with a low risk of bias provided consistent, generalisable results for the outcome. The overall quality of evidence was downgraded by one level if one of the factors described above was not met. Likewise, if two or three factors were not met, then it was downgraded by two or three levels, respectively. Thus, the GRADE approach resulted in four levels of quality of evidence: high, moderate, low, and very low. In case of only one study measuring an outcome, data were considered to be sparse and inconsistent and the evidence was labelled as ’low quality evidence’.

Results Study selection

The computer generated search resulted in 2654 references in Pubmed, 404 in EMBASE, 62 in PsychINFO, 206 in the Cochrane Central Register of Controlled Trials, 23 in the Cochrane Occupational Health Field. After exclusion of the duplicated references, both reviewers (MTD and KIP) read 3067 titles and abstracts. Disagreements were resolved in a consensus meeting. The most important reasons for exclusion were: the study design was not an RCT, the study population consisted of sick listed workers, and outcome measure was not LBP or NP incidence or intensity. Hand searching of the reference lists of relevant review articles did not result in any new articles. Finally, 10 studies were included in this systematic review (figure 1). Risk of bias assessment

Before contacting the authors, 52 risk of bias criteria were scored ‘don’t know’. After authors had provided additional information, 16 risk of bias criteria were still scored ‘don’t know’. Table 1 shows the risk of bias assessment scores for the included studies. Seven studies were classified as ‘low risk of bias’ and three as ‘high risk of bias’. Few studies were able to keep the participants blinded for the intervention (criterion C), and only one study was able to successfully blind the care provider (criterion D). Some studies did not report at all or reported insufficiently on these criteria. Blinding in workplace settings is not really possible21, so there is always a potential risk of bias in this field. No study blinded the outcome assessor (criterion E) seeing that self-reported subjective experience of pain was the outcome. Further, most studies did not report on the use of co-interventions or compliance with the intervention.

26

Chapter 2

Figure 1. Flow chart of selection process.

Electronic search of 5 databases (n=3349 ): Pubmed (n=2654 ), EMBASE (n=404), PsychINFO (n=62), CENTRAL (n=206), Cochrane Occupational Health Field (n=23) Studies excluded due to doublings (n=282) Potentially relevant studies identified and screened for retrieval (n=3067) Studies excluded based on abstract : inclusion criteria were not met (n=3036) Studies retrieved for more detailed evaluation (n=31) Studies excluded (n=21) Studies included in the systematic review (n=10)

Table 1. Characteristics of included studies.

Baseline

Co-interventions

Compliance

Timing

Total score

L

Selective report

K

Intention to treat

J

Drop-out

I

Outcome blinded

H

Care provider blinded

G

Patient blinded

F

Concealment

Haukka et al.24 Lengsfeld et al.28 Brisson et al.22 Heuvel et al.26 Gerr et al.30 Rempel et al.31 Rempel et al.25 Conlon et al.27 Cook et al.23 Mekhora et al.29

E

Randomisation

Table 1. Characteristics of included studies. Criterion A B C D

1 1* 1* 1 1 1 1* 1 1 ?

1 1* ? 1* ? 1* 1* 1* 0* ?

0* 1 0* 0* ? 0* 0* 0* 0* 0

0* 1 0* 0* ? 0* 0* 0* 0* ?

0 0 0 0 0 0 0 0 0 0

1 1 1 1 1 0 0 0 1 1

1 1 1* 1* 1 1 1 1 1 ?

1 1 1 0 1 1 1 1 1 1

1 1 1 0 1 1 1 0 0 ?

1* 0* 1* 1* ? ? ? ? 0* ?

1 0* 0* 1* 0 0* 0* ? 0* ?

1 1 1 1 1 1 1 1 1 1

9 9 7 7 6 6 6 5 5 3

* Before contacting the first authors, items were scored “don’t know”.

Chapter 2

27

Study characteristics

Table 2 shows the characteristics of the studies included. The number of participants varied from 59 to 627.22;23 All studies, except two, were conducted in an office environment.24;25 Nine interventions were classified as physical ergonomic interventions, and one as an organisational ergonomic intervention.26 The study of Haukka et al. (2008) was classified as a physical ergonomic intervention, because the participatory ergonomic programme predominantly resulted in adjustments to the workplace or new equipment.24 Five studies were conducted on workers with and without symptoms22-25;27, three on workers with symptoms26;28;29, and two on workers without symptoms.30;31 The duration of follow-up among studies varied from 6 weeks23 to 2 years.24-28 One study reported on LBP only28, five studies on NP only25-27;30;31, and four studies reported on both LBP and NP.22-24;29 Measurements used to determine incidence/prevalence among studies varied, and included using a manikin to identify the body region24, the use of medication for symptoms, cut-off points on self-reported discomfort or pain intensity scales, and/or subsequent diagnosis by a healthcare provider.22;27;30;31 Measurements on pain intensity also varied, and some studies used a visual analogue scale (VAS)26;28;31, a 5-point Likert scale25, or a 10-point discomfort scale.29 One high risk of bias study (n = 85) showed that a physical ergonomic intervention (i.e. workstation intervention) was not effective in reducing LBP and NP intensity in the short term.29 However, the study only performed within-group comparisons and did not perform any between-group comparisons. The authors did not respond to our request to provide additional information on between-group comparisons, this study was excluded from the analyses of this review.

28

Chapter 2

Chapter 2

29

Participatory ergonomic programme: on the basis of active group work, workers identified problems, evaluated changes, and implemented them in collaboration with management and technical staff 1. Curved seat pan chair plus miscellaneous items 2. Flat seat pan chair plus miscellaneous items

Computer software stimulating 5-minutes computer breaks after 35 minutes computer usage. 7 seconds break after using the computer for 5 minutes, a booklet and a neck and upperlimb disorder risk test

59 call centre workers with and without LBP and NP at baseline

504 kitchen workers with and without LBP and NP at baseline

480 garment workers with and without neck/shoulder pain at baseline

268 office employees with neck/shoulder pain at baseline

Cook et al.(2004), 23 Australia

Haukka et al.(2008), 24 Finland

Rempel et al.(2007), 25 USA

Van den Heuvel et al.(2003), 26 Netherlands

Adjustments to the desk surface to support the fore-arm, and adjustments on keyboard and mouse position

Ergonomic training: two sessions of 3 hours on workplace adjustments workplace (postural and visual components), and work organisation

627 university workers with and without LBP and NP at baseline

Brisson et al.(1999), 22 Canada

Intervention(s)

Participants randomised

Table 2. Characteristics of included studies.

Study, country

8 weeks

4 months

Miscellaneous items: footrest, storage box, side table, task lamp, and reading glasses

Booklet with information on neck and upper limb disorders and a neck and upper-limb disorder risk test

Prevalence of LBP and NP (manikin illustration)

24 months

No participatory ergonomics programme

Neck/shoulder pain severity (VAS 1-10)

Pain intensity in the past month (5 point scale: 1 “little painful”- 5 “very painful”)

Neck/shoulder pain intensity

(Nordic questionnaire)

Presence of LBP and NP

Having pain 3 days or more in past 7 days greater than >5 on a VAS (0-10) were referred to an occupational therapist for diagnosis. Prevalence is determined by a positive diagnosis

Prevalence of LBP and NP

6 weeks

6 months

No ergonomic training

Outcome

Workplace adjustments according to Australian standards

Follow-up duration

Control intervention

NP: 3.00 (SD 2.33)* vs. 3.14 (SD 2.52); NS

1. Difference in slope (interventioncontrol) of NP score for change over time = -0.34 (95% CI -0.41 to -0.28)**; p= not reported 2. Difference in slope (interventioncontrol) of NP score for change over time: -0.14 (95% CI -0.22 to -0.07)**; p= not reported

a. LBP: 126/263 vs. 111/241†; NS b. LBP: 145/241 vs. 135/241†; NS a. NP: 176/263 vs.174/241†; NS b. NP: 184/263 vs. 159/241†; NS

LBP: 4/30 vs. 8/29†; NS NP: 5/30 vs. 8/29†; NS

LBP: 22/283 vs. 24/339*; NS NP: 36/282 vs. 46/341*; NS

Results intervention vs. control

7

6

9

5

Risk of bias 7

30

Chapter 2

6 months

1. Alternate intervention: postural intervention with workplace changes based on results from a prospective study on MSD 2. Conventional intervention: postural intervention with workplace change based on OSHA, NIOSH and private industry rules

No intervention: continue keying in usual posture and no workstation changes

356 computer workers without neck/shoulder pain at baseline

Gerr et al.(2005), 30 USA

Cross-over design after 14 weeks, intervention becomes own control

Use of unadjusted workstation

Workstation intervention: Using a software program, advices on computer workstation adjustments were given on monitor and keyboard height and the use of foot stools and document holders

85 office workers with tension neck symptoms at baseline

Mekhora et al.(2000), 29 Thailand

24 months

Office chair without micro rotation function

Office chair with micro rotation function underneath the seat to prevent long term sitting

280 office workers with chronic recurrent LBP at baseline

Lengsfeld et al.(2007), 28 Germany

Follow-up duration 12 months

1. Alternative mouse: vertical handle for grapping, flat base to support ulnar side of the hand and a roller ball for tracking 2. Fore arm board plus conventional mouse 3. Fore arm board plus alternative mouse

206 (supportive) engineering staff with and without neck/shoulder pain at baseline

Conlon et al.(2008), 27 USA

Control intervention Conventional mouse: hand in almost fully pronated posture and an optical LED for tracking

Intervention(s)

Participants randomised

Study, country

Table 2. Continued.

Discomfort score of >=6 (0-10 discomfort scale) or medication use for musculoskeletal discomfort on any day of the week

Incidence of neck/shoulder pain

(VAS 0-10, “no pain-extreme pain”)

Discomfort of LBP and NP

(VAS 0-100mm)

Lumbar Pain score

Workers with discomfort rates of >5 (0-10 scale) in the past 7 days or used medication for upper body discomfort were referred to an occupational physician for diagnosis. Incidence of NP is a positive diagnosis

Incidence of neck/shoulder pain

Outcome

NP: 38/114 vs. 33/109; NS NP: 36/116 vs. 33/109; NS

LBP: change score from baseline: -1.436; NS (within group comparison) NP: change score from baseline: -1.197; NS (within group comparison)

a. LBP: 39.75 (SD 17.79) vs. 38.53 (SD 17.46); NS b. LBP: 34.56 (SD 18.41) vs. 33.32 (SD 19.39); NS

1. NP: 4/52 vs. 3/52; NS 2. NP: 8/51 vs. 3/52; NS 3. NP: 3/51 vs. 3/52; NS

Results intervention vs. control

6

3

9

Risk of bias 5

Chapter 2

31

1. Ergonomic training and trackball 2. Ergonomic training plus arm board 3. Ergonomic training plus trackball plus arm board

182 customer service operators without neck/shoulder pain at baseline

Rempel et al.(2006), 31 USA

Ergonomic training: erect sitting posture, adjustments on the height of the chair, arm supports, work surface, monitor, and adjustments of mouse and keyboard location

Control intervention 12 months

Follow-up duration

(VAS 0-10)

Neck/shoulder pain intensity

Pain intensity scores >5 (VAS 010) OR pain medication for two days or more per week due to neck/shoulder/upper extremity complaints) were referred to occupational physician for diagnosis. Incidence of NP is a positive diagnosis

Incidence of neck/shoulder pain

Outcome

1. NP: 2.2 (SD 2.2)* vs. 1.8 (SD 1.9)* 2. NP: 2.6 (SD 2.8)* vs. 1.8 (SD 1.9)* 3. NP: 2.0 (SD 2.4)* vs. 1.8 (SD 1.9)*

1. NP : 6/35 vs. 19/43 2. NP : 6/40 vs. 19/43 3. NP : 8/40 vs. 19/43

Results intervention vs. control

Abbreviations: LBP, low back pain; NP, neck pain; MSD, musculoskeletal disorders; SD, standard deviation; NS, not significant; VAS, visual analogue scale. * information not reported in study and provided by first authors on request. ** unadjusted estimates. a. results at short term follow-up. b. results at final follow-up measurement. †: frequencies derived from percentages.

Intervention(s)

Participants randomised

Study, country

Table 2. Continued. Risk of bias 6

LBP incidence/prevalence

Short term Two studies with low risk of bias (total n = 1131)22;24 and one study with high risk of bias (n = 59) evaluated the effectiveness of an ergonomic intervention on LBP prevalence. The participants included in these studies consisted of workers with or without LBP at baseline. The physical ergonomic interventions included an ergonomic training incorporating workplace adjustments for university employees22, a participatory ergonomics programme instituting workplace changes for kitchen workers24, and computer workplace adjustments for call centre workers.23 The quality on LBP prevalence was downgraded with two levels. The results were inconsistent because in one study the LBP prevalence decreased23, while in the two other studies the LBP prevalence remained the same.22;24 The results of the pooled data were indirect, because the effect was largely determined by the high weight (87.7%) in the meta-analysis of one study conducted on kitchen workers.24 Therefore, there is low quality evidence from three studies (N = 1190) that there is no statistically significant difference in the reduction in LBP prevalence in the short term (RR 1.03; 95% CI 0.86-1.22) between groups that received a physical ergonomic intervention compared to groups receiving no such intervention (figure 2a). Long term One low risk of bias study (N=504) evaluated the effectiveness of a physical ergonomic intervention on LBP prevalence in the long term among kitchen workers with and without LBP at baseline. A participatory ergonomic programme was no more effective than no intervention on 2-year prevalence of LBP.24 There is low quality evidence that a physical ergonomic intervention is no more effective than no such intervention at reducing LBP prevalence in the long term. NP incidence/prevalence

Short term Three low risk of bias studies (total n = 1487)22;24;30 and one high risk of bias study (n = 59)23 compared the effectiveness of a physical ergonomic intervention to no intervention on NP incidence/prevalence. The study of Gerr et al. (2005) was evaluated as regards NP free workers30, while the three other studies included workers with and without NP.22-24 Ergonomic interventions included ergonomic training incorporating workplace adjustments for university employees22, an alternate or conventional postural intervention with workstation changes for computer workers30, computer workplace adjustments for call centre workers23, and a participatory ergonomic programme consisting of workplace changes for kitchen

32

Chapter 2

workers.24 The quality of evidence on this outcome was downgraded with one level. The results were indirect, because the pooled effect was largely determined by the high weight (84.7%) in the meta-analysis of one study that was conducted on kitchen workers.24 Therefore, there is moderate quality evidence from four studies (n = 1546) that there is no statistically significant difference in the reduction of NP incidence/prevalence at the short term (RR 0.93; 95% CI 0.84-1.03) between groups that received a physical ergonomic intervention compared to groups receiving no such intervention (figure 2b). Long term Two RCTs with low risk of bias (n = 686)24;31 and one high risk of bias RCT (n = 206)27 were identified. All the interventions under study were classified as physical ergonomic interventions and were conducted on workers with and without NP at baseline. Rempel et al. (2006) compared the effectiveness of three ergonomic interventions among customer service operators (ergonomic training and trackball, ergonomic training plus arm board and ergonomic training plus trackball plus arm board) to ergonomic training31, and found that an ergonomic training plus an arm board, even when combined with a trackball, was significantly more effective than the ergonomic training only. Among engineering staff, Conlon et al. (2008), however, did not find any significant differences when an alternative mouse; an arm board combined with an alternative mouse or an arm board with a conventional mouse were compared to a conventional mouse.27 Haukka et al. (2008) showed that a participatory ergonomic programme was no more effective than no intervention among kitchen workers regarding 2-year prevalence of NP.24 The quality of evidence on this outcome was downgraded with two levels. Results were inconsistent and pooled data were imprecise, meaning that the width of the confidence interval of the pooled data made it impossible to support or refute the effectiveness of physical ergonomic interventions. Therefore, there is low quality evidence from three studies (n = 892) that there is no statistically significant difference in the reduction of NP incidence/prevalence at the long term (RR 0.79; 95% CI 0.41-1.53) between groups that received a physical ergonomic intervention compared to groups that received no such intervention (figure 2c).

Chapter 2

33

Figure 2a. Meta-analyses of three studies on physical and organisational ergonomic interventions compared to a control intervention in the reduction of short term LBP prevalence.

Figure 2b. Meta-analyses of four studies on physical and organisational ergonomic interventions compared to a control intervention in the reduction of short term NP incidence/prevalence.

Figure 2c. Meta-analyses of three studies on physical and organisational ergonomic interventions compared to a control intervention in the reduction of long term NP incidence/prevalence.

34

Chapter 2

LBP intensity

Short and long term One low risk of bias study (N = 157) investigated the effects of a physical ergonomic intervention on the reduction in LBP intensity at the short and long term. Using a 2-year follow-up period, Lengsfeld et al. (2007) showed that a new office chair, with an electric motor underneath the seat to prevent prolonged sitting, was no more effective than the same chair without an electric motor.28 There is low quality evidence that a physical ergonomic intervention is no more effective than no such interventions in reducing LBP intensity at both the short and long term. NP intensity

Short term Two low risk of bias studies (total n = 748) evaluated an ergonomic intervention.25;26 One study investigated the effectiveness of an organisational ergonomic intervention among office workers, but found that rest breaks were no more effective than an informative brochure to reduce NP intensity.26 A study of garment workers evaluated two physical ergonomic interventions: a chair with a curved seat and miscellaneous items, and a chair with a flat seat and miscellaneous items. Compared to a group with miscellaneous items (e.g. a footrest, storage box, side table, task lamp and reading glasses), both chairs were significantly more effective in reducing NP intensity.25 The garment study showed, on a 5-point Likert scale, that the curved seat pan chair reduced NP intensity by 0.34 points, while the flat seat pan chair reduced NP intensity by 0.14 points. It should be noted that these significant results were found in a subgroup of 277 workers with NP at baseline, while a total 480 workers were randomised to one of the three groups. The garment study did not describe the intervention effects among the excluded subgroup (n = 203, without NP at baseline) or among the entire study population (n = 480). The use of two different pain scales (continuous and categorical) and the use of different types of ergonomic interventions among the studies made a meta-analysis on this outcome impossible. In summary, there is low quality evidence from one study (n = 268)26 that an organisational ergonomic intervention is no more effective than no such intervention in reducing NP intensity in the short term. Based on the significant reduction in NP intensity found in the garment study (n = 277)25, there is low quality evidence that a physical ergonomic intervention (i.e. curved and flat seat pan chair) is significantly more effective for reducing NP intensity in the short term than no ergonomic intervention. Long term One study with a low risk of bias (n = 182) evaluated the effectiveness of three physical

Chapter 2

35

ergonomic interventions among customer service operators (ergonomic training and trackball, ergonomic training plus arm board and ergonomic training plus trackball plus arm board) to no ergonomic training.31 Ergonomic interventions that combined the use of an arm board support and an ergonomic training were significantly more effective in reducing NP intensity than ergonomic training only. As regards the use of a trackball, no significant effects were reported on NP intensity. Based on the significant reduction in NP intensity found in this single study, there is low quality evidence that a physical ergonomic intervention (i.e. arm board support) is significantly more effective in reducing NP intensity in the long term than no ergonomic intervention.

Discussion This review investigated the effectiveness of physical and organisational ergonomic interventions on the prevention and reduction in LBP and NP among non-sick listed workers. The findings of this review showed that there is low to moderate evidence that ergonomic interventions were no more effective than control interventions on short and long term LBP and NP incidence/prevalence, LBP intensity and short term NP intensity. However, we found low quality evidence that in the short term a physical ergonomic intervention (i.e. curved and flat seat pan chair) was significantly more effective in reducing NP intensity than no ergonomic intervention. There was also low quality evidence that in the long term a physical ergonomic intervention (i.e. arm board) was significantly more effective in reducing NP intensity than no ergonomic intervention. However, these findings were obtained from two studies only.25;31 The results of the current review have to be interpreted with caution because of the limited number of studies per outcome and the heterogeneity in populations (symptomatic and non symptomatic), interventions, controls and outcomes. The generalisability of the results to the entire working population is low, because populations studied only consisted of office workers, garment workers, and kitchen workers. Further, the results of the pooled data on short-term prevention of LBP and NP were dominated by a large study that was conducted among kitchen workers.24 Moreover, almost all other studies were on NP and were conducted in an office setting evaluating physical ergonomic interventions (i.e. workstation adjustments). At present, RCTs on organisational ergonomic interventions to prevent and reduce LBP and NP are lacking. Despite the limited number of included RCTs, this review provides solid epidemiological evidence of the effectiveness of ergonomic interventions on LBP and NP. Findings compared to other reviews

The conclusions of the current review differ somewhat compared from those of other reviews.5;13-15 Compared to previous reviews, this one specifically focussed on LBP or NP,

36

Chapter 2

while others included a larger variety of symptoms (i.e. neck/upper extremity pain, upper extremity musculoskeletal disorders or visual symptoms). Further, this review excluded study designs other than RCTs and excluded individual worker interventions. Moreover, none of the other reviews performed a meta-analysis and none used the GRADE classification system for levels of evidence. Explanation of the findings

A number of factors may explain the results found in most studies. Due to the small sample sizes, there is a lack of power to detect positive effects. A meta-analysis was conducted that increases the power, but the results of the meta-analysis showed no statistically significant differences in effect. Six studies used relatively short follow-up periods that varied from 6 weeks to 6 months and found no effect. This might indicate that follow-up periods shorter than 6 months are too short to measure an effect. Furthermore, longer follow-up periods make it possible to measure intervention sustainability16 and enable identification of delayed intervention effects. More measurements during follow-up may also be needed as LBP and NP are both marked by periods of remission and exacerbation.32;33 By using one or two follow-up points only, the incidence/prevalence of LBP and NP may be over-or underestimated. Therefore, more advanced study designs and statistical methods are recommended, for example study designs with repeated measurements.34 Furthermore, a considerable number of studies in this review included both workers with and without symptoms at baseline, and as a consequence may suffer from prevalence-incidence bias. Symptomatic workers at baseline may recover during follow-up, while workers without symptoms at baseline may in time develop LBP and NP. Further, because baseline pain intensity scores were low, little room was left for improvement on pain intensity scores.35;36 Another reason that no effect was found may be related to the exposure to occupational risk factors for LBP and NP. In their conceptual model, Westgaard and Winkel (1997) hypothesised that the implementation of an ergonomic intervention may change the workers’ mechanical exposure and/or may affect the physical or psychosocial risk factors for musculoskeletal health, which in turn would lead to improved outcomes on musculoskeletal health.16 In the current review, eight out of 10 studies were conducted among office workers. All but two26;28 ergonomic interventions were aimed at optimising the workers’ mechanical workload which in turn would reduce the physical risk factors for NP. The most important physical risk factor for NP and upper limb symptoms is repetitiveness combined with forceful exertions.37-39 However, the exposure to such a physical load among office workers is very small. Psychosocial factors may also play a role in the onset of NP among office workers5, however, none of the ergonomic interventions were targeted at the psychosocial workload. Another possibility is that the ergonomic interventions did not target the most important risk factors. However, the issue of risk factors for LBP and NP is still poorly understood, particularly which risk factors are most likely to change through ergonomic interventions. In

Chapter 2

37

addition, risk factors outside the workplace may not be affected by ergonomic interventions.40 Despite the fact that an RCT should control for unforeseen factors, according to some researchers, the work life environment may be too complex for such control. Although we agree that other study designs can add to our knowledge of the mechanisms of ergonomic interventions, in our opinion the RCT design is the gold standard to evaluate the effectiveness of ergonomic interventions. The view that RCTs are only applicable in occupational health settings and to ergonomic interventions is debatable because contamination between workers in the intervention and control groups can easily occur. To avoid contamination, randomisation at the workplace level (department or firm) is recommended.41 In our review, only two studies performed a so-called cluster randomisation procedure.24;26 Finally, it may be that workers were not compliant with the ergonomic intervention. An intervention may be perfectly designed, but high compliance is still very important for its effectiveness.42 From the scoring of the methodological quality criteria, it appeared that most studies had either insufficient levels of compliance or did not report on compliance at all. Reporting on this criterion is, therefore, strongly recommended. To increase worker’s compliance, the use of an appropriate implementation strategy may be beneficial.43 For instance, among floor layers an adequate implementation strategy was effective in reducing severe knee problems.44 Furthermore, to improve interventions, authors mentioned that the combination of quantitative studies with qualitative studies would be worthwhile in order to examine participant’s experiences with the intervention and the intervention effects on different subgroups and settings.21 Subsequently, the new insights into the working mechanism of an intervention can be used for the development of new ergonomic interventions.

Strengths and limitations of the review

One of the main strengths of this review is that we only included RCTs, which are the studies least susceptible to bias. Furthermore, this review performed a meta-analysis on the results of the ergonomic interventions. The present review has some limitations. The aim of this review was to summarise the existing knowledge and evidence concerning the effectiveness of ergonomic interventions on LBP and NP. A systematic review is a form of observational research and, therefore, selection bias may have occurred. Even though a highly sensitive literature search was conducted, it is still possible that studies were missed in this review. Three studies evaluated more than one ergonomic intervention. To avoid double counting of these studies, we chose to include the most effective intervention from these studies in the meta-analyses. This may have influenced the results, leading to an overestimation of the intervention effect. If studies did not report risk ratios, we calculated them using uncorrected study data. This also may have led to an overestimation of the effect size. However, because we did not find a statistically significant difference in effectiveness, these biases can be excluded.

38

Chapter 2

Conclusion

This review showed low to moderate quality evidence that physical and organisational ergonomic interventions were not more effective on short and long term LBP and NP incidence/prevalence, on short and long term LBP intensity than no ergonomic intervention. In the short term, a physical ergonomic intervention (i.e. curved and flat seat pan chair) was significantly more effective in reducing NP intensity than no ergonomic intervention. There was also low quality evidence that in the long term a physical ergonomic intervention (i.e. arm board) was significantly more effective in reducing NP intensity than no ergonomic intervention. However, these findings were obtained from two studies only. In conclusion, ergonomic interventions were usually not effective in preventing or reducing LBP and NP among non-sick listed workers.

Chapter 2

39

Reference list 1. Borghouts JA, Koes BW, Vondeling H, and Bouter LM. Cost-of-illness of neck pain in The Netherlands in

1996. Pain 1999;80:629-36.

2. van Tulder MW, Koes BW, and Bouter LM. A cost-of-illness study of back pain in The Netherlands. Pain

1995;62:233-40.

3. Bongers PM, Ijmker S, van den Heuvel S, and Blatter BM. Epidemiology of work related neck and

upper limb problems: psychosocial and personal risk factors (part I) and effective interventions from a



bio behavioural perspective (part II). J Occup Rehabil. 2006;16:279-302.

4. van Tulder M, Koes B, and Bombardier C. Low back pain. Best Pract.Res.Clin.Rheumatol. 2002;16:761-75. 5. Linton SJ and van Tulder MW. Preventive interventions for back and neck pain problems: what is the

evidence? Spine 2001;26:778-87.

6. van Dieën JH, van der Beek AJ. Work-Related Low-Back Pain: Biomechanical Factors and Primary

Prevention. Kumar S, ed. Ergonomics for Rehabilitation Professionals. FL: Boca Raton, 2009:359-95.

7. Bigos SJ, Holland J, Holland C, Webster JS, Battie M, and Malmgren JA. High-quality controlled trials on

preventing episodes of back problems: systematic literature review in working-age adults. Spine J 2009;9:147-68.

8. Tveito TH, Hysing M, and Eriksen HR. Low back pain interventions at the workplace: a systematic

literature review. Occup Med.(Lond) 2004;54:3-13.

9. van Poppel MNM, Hooftman WE, and Koes BW. An update of a systematic review of controlled clinical

trials on the primary prevention of back pain at the workplace. Occup Med.(Lond) 2004;54:345-52.

10. Ammendolia C, Kerr MS, and Bombardier C. Back belt use for prevention of occupational low back

pain: a systematic review. J Manipulative Physiol Ther. 2005;28:128-34.

11. Martimo KP, Verbeek J, Karppinen J et al. Effect of training and lifting equipment for preventing back

pain in lifting and handling: systematic review. BMJ 2008;336:429-31.

12. van Duijvenbode ICD, Jellema P, van Poppel MNM, and van Tulder MW. Lumbar supports for prevention

and treatment of low back pain. Cochrane Database Syst.Rev. 2008;CD001823.

13. Boocock MG, McNair PJ, Larmer PJ et al. Interventions for the prevention and management of neck/

upper extremity musculoskeletal conditions: a systematic review. Occup Environ Med. 2007;64:291-303.

14. Brewer S, Van Eerd D, Amick BC et al. Workplace interventions to prevent musculoskeletal and visual

symptoms and disorders among computer users: a systematic review. J Occup Rehabil. 2006;16:325-58.

15. Burton AK, Balague F, Cardon G et al. Chapter 2. European guidelines for prevention in low back pain:

November 2004. Eur.Spine J 2006;15 Suppl 2:S136-S168.

16. Westgaard R.H. and Winkel J. Ergonomic intervention research for improved musculoskeletal health:

A critical review. Int.J.Ind.Ergon. 1997;20:463-500.

17. Furlan AD, Pennick V, Bombardier C, and van Tulder M. 2009 updated method guidelines for systematic

reviews in the Cochrane Back Review Group. Spine (Phila Pa 1976.) 2009;34:1929-41.

18. Verbeek J, Salmi J, Pasternack I et al. A search strategy for occupational health intervention studies.

40

Occup Environ Med. 2005;62:682-7.

Chapter 2

19. Atkins D, Best D, Briss PA et al. Grading quality of evidence and strength of recommendations. BMJ

2004;328:1490.

20. Guyatt GH, Oxman AD, Kunz R et al. Going from evidence to recommendations. BMJ 2008;336:1049-51. 21. Schonstein E and Verbeek JH. Occupational health systematic reviews: An overview. Work 2006;26:255-8. 22. Brisson C, Montreuil S, and Punnett L. Effects of an ergonomic training program on workers with video

display units. Scand.J Work Environ Health 1999;25:255-63.

23. Cook C and Burgess-Limerick R. The effect of forearm support on musculoskeletal discomfort during call

centre work. Appl.Ergon 2004;35:337-42.

24. Haukka E, Leino-Arjas P, Viikari-Juntura E et al. A randomised controlled trial on whether a participatory

ergonomic intervention could prevent muskuloskeletal disorders. Occup Environ Med. 2008;65:849-956.

25. Rempel DM, Wang PC, Janowitz I, Harrison RJ, Yu F, and Ritz BR. A randomized controlled trial

evaluating the effects of new task chairs on shoulder and neck pain among sewing machine operators:



the Los Angeles garment study. Spine 2007;32:931-8.

26. van den Heuvel SG, de Looze MP, Hildebrandt VH, and The KH. Effects of software programs stimulating

regular breaks and exercises on work-related neck and upper-limb disorders. Scand.J Work Environ



Health 2003;29:106-16.

27. Conlon CF, Krause N, and Rempel DM. A randomised controlled trial evaluating an alternative mouse

and forearm support on upper body discomfort and musculoskeletal disorders among engineers. Occup



Environ Med. 2008;65:311-8.

28. Lengsfeld M, Konig IR, Schmelter J, and Ziegler A. Passive rotary dynamic sitting at the workplace by

office-workers with lumbar pain: a randomized multicenter study. Spine J 2007;7:531-40.

29. Mekhora K, Liston CB, Nanthavanij S, and Cole JH. The effect of ergonomic intervention on discomfort

in computer users with tension neck syndrome. Int.J.Ind.Ergon. 2000;26:367-79.

30. Gerr F, Marcus M, Monteilh C, Hannan L, Ortiz D, and Kleinbaum D. A randomised controlled trial of

postural interventions for prevention of musculoskeletal symptoms among computer users. Occup Environ



Med. 2005;62:478-87.

31. Rempel DM, Krause N, Goldberg R, Benner D, Hudes M, and Goldner GU. A randomised controlled

trial evaluating the effects of two workstation interventions on upper body pain and incident



musculoskeletal disorders among computer operators. Occup Environ Med. 2006;63:300-6.

32. Cassidy JD, Côté P, Carroll LJ, and Kristman V. Incidence and course of low back pain episodes in the

general population. Spine 2005;30:2817-23.

33. Côté P, Cassidy JD, Carroll LJ, and Kristman V. The annual incidence and course of neck pain in the

general population: a population-based cohort study. Pain 2004;112:267-73.

34. de Vet HC, Heymans MW, Dunn KM et al. Episodes of low back pain: a proposal for uniform definitions

to be used in research. Spine 2002;27:2409-16.

35. Kovacs FM, Abraira V, Royuela A et al. Minimal clinically important change for pain intensity and

disability in patients with nonspecific low back pain. Spine 2007;32:2915-20.

Chapter 2

41

36. van der Roer N, Ostelo RWJG, Bekkering GE, van Tulder MW, and de Vet HCW. Minimal clinically

important change for pain intensity, functional status, and general health status in patients with



nonspecific low back pain. Spine 2006;31:578-82.

37. Malchaire J, Cock N, and Vergracht S. Review of the factors associated with musculoskeletal problems

in epidemiological studies. Int Arch.Occup Environ Health 2001;74:79-90.

38. National Research Council & Institute of Medicine. Low back and upper extremities. Washington, DC:

National Academy Press, 2001.

39. van der Windt DA, Thomas E, Pope DP et al. Occupational risk factors for shoulder pain: a systematic

review. Occup Environ Med. 2000;57:433-42.

40. Frank JW, Kerr MS, Brooker AS et al. Disability resulting from occupational low back pain. Part I: What

do we know about primary prevention? A review of the scientific evidence on prevention before disability



begins. Spine 1996;21:2908-17.

41. Verbeek J. The occupational health field in the cochrane collaboration. Ind Health 2007;45:8-12. 42. Cleemput I and Kesteloot K. Economic implications of non-compliance in health care. Lancet

2002;359:2129-30.

43. van der Molen HF, Sluiter JK, Hulshof CT, Vink P, and Frings-Dresen MH. Effectiveness of measures and

implementation strategies in reducing physical work demands due to manual handling at work.



Scand.J.Work Environ.Health 2005;31 Suppl 2:75-87.

44. Jensen LK and Friche C. Effects of training to implement new tools and working methods to reduce knee

42

load in floor layers. Appl.Ergon. 2006;65:20-7.

Chapter 2